YOLOv5修改注意力机制CBAM

直接上干货

CBAM注意力机制是由通道注意力机制（channel）和空间注意力机制（spatial）组成。

传统基于卷积神经网络的注意力机制更多的是关注对通道域的分析，局限于考虑特征图通道之间的作用关系。CBAM从 channel 和 spatial 两个作用域出发，引入空间注意力和通道注意力两个分析维度，实现从通道到空间的顺序注意力结构。空间注意力可使神经网络更加关注图像中对分类起决定作用的像素区域而忽略无关紧要的区域，通道注意力则用于处理特征图通道的分配关系，同时对两个维度进行注意力分配增强了注意力机制对模型性能的提升效果。

CBAM中的通道注意力机制模块流程图如下。先将输入特征图分别进行全局最大池化和全局平均池化，对特征映射基于两个维度压缩，获得两张不同维度的特征描述。池化后的特征图共用一个多层感知器网络，先通过一个全连接层下降通道数，再通过另一个全连接恢复通道数。将两张特征图在通道维度堆叠，经过 sigmoid 激活函数将特征图的每个通道的权重归一化到0-1之间。将归一化后的权重和输入特征图相乘。

yaml 配置文件如下

# YOLOv5 🚀 by Ultralytics, GPL-3.0 license

# Parameters
nc: 6  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.25  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, CBAM, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 10
  ]

# YOLOv5 v6.0 head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 14

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 18 (P3/8-small)

   [-1, 1, Conv, [256, 3, 2]],
   [[-1, 15], 1, Concat, [1]],  # cat head P4
   [-1, 3, C3, [512, False]],  # 21 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 11], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 24 (P5/32-large)

   [[18, 21, 24], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

common加入以下代码

# CBAM
class ChannelAttention(nn.Module):
    def __init__(self, in_planes, ratio=16):
        super(ChannelAttention, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.max_pool = nn.AdaptiveMaxPool2d(1)
        self.f1 = nn.Conv2d(in_planes, in_planes // ratio, 1, bias=False)
        self.relu = nn.ReLU()
        self.f2 = nn.Conv2d(in_planes // ratio, in_planes, 1, bias=False)
        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        avg_out = self.f2(self.relu(self.f1(self.avg_pool(x))))
        max_out = self.f2(self.relu(self.f1(self.max_pool(x))))
        out = self.sigmoid(avg_out + max_out)
        return out


class SpatialAttention(nn.Module):
    def __init__(self, kernel_size=7):
        super(SpatialAttention, self).__init__()
        assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
        padding = 3 if kernel_size == 7 else 1
        # (特征图的大小-算子的size+2*padding)/步长+1
        self.conv = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=False)
        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        # 1*h*w
        avg_out = torch.mean(x, dim=1, keepdim=True)
        max_out, _ = torch.max(x, dim=1, keepdim=True)
        x = torch.cat([avg_out, max_out], dim=1)
        # 2*h*w
        x = self.conv(x)
        # 1*h*w
        return self.sigmoid(x)


class CBAM(nn.Module):
    # CSP Bottleneck with 3 convolutions
    def __init__(self, c1, c2, ratio=16, kernel_size=7):  # ch_in, ch_out, number, shortcut, groups, expansion
        super(CBAM, self).__init__()
        self.channel_attention = ChannelAttention(c1, ratio)
        self.spatial_attention = SpatialAttention(kernel_size)

    def forward(self, x):
        out = self.channel_attention(x) * x
        # c*h*w
        # c*h*w * 1*h*w
        out = self.spatial_attention(out) * out
        return out

YOLO 的